Article ID: 951434 - View products that this article applies to.
In Windows Server 2008, when you run the cluster validation process on a cluster configuration that is using iSCSI for storage, the Validate Network Communication test may fail on the iSCSI networks. Additionally, you may receive the following error message:
Network interfaces interface 1 and interface 2 are on the same cluster network, yet either address address 2 is not reachable from address address 1 or the ping latency is greater than the maximum allowed 500 milliseconds.
This problem occurs if a multipath iSCSI network is configured as follows:
Node A, Path 1 = 192.168.1.1
Node A, Path 2 = 192.168.2.1
Node B, Path 1 = 192.168.1.2
Node B, Path 2 = 192.168.2.2
Path 1 on both nodes uses controller 1 on the storage.
Path 2 on both nodes uses controller 2 on the storage.
In this scenario, Path 1 uses the same subnet for Node A and for Node B. Also, Path 2 uses the same subnet for Node A and for Node B.
Disjoint networks, such as networks that are directly attached to the storage, should not use the same subnet. Because the storage is attached directly, packets that are sent from Path 1 on Node A will not reach Path 1 on Node B. Also, packets that are sent from Path 2 on Node A will not reach Path 2 on Node B because both networks are disjoint networks. The cluster validation test fails because the disjoint networks are using the same subnet.
To resolve this problem, examine the cluster validation logs and the cluster configuration to determine whether the scenario that is discussed in the "Cause" section applies to you. If this scenario does apply, change the network configuration so that disjoint networks are not located on the same subnet. Read the documentation for your storage system. Or, contact the storage vendor to determine how to make this change correctly.
Article ID: 951434 - Last Review: September 11, 2010 - Revision: 1.1